Quote:
Originally Posted by Taco MacArthur
Chris Osgood was the definition of "league average goaltender":
http://hockeygoalies.org/bio/osgood.html
He was marginally better than that in his first Detroit stint: +46.6 goals above average between 199394 and 200001 (and a 1.9 in the playoffs over that span), but unless you've got a peculiar definition of "generation", that's nowhere near top five in Osgood's case.
We can argue about inadequacies of save percentage as a metric, but any biases there aren't going to push him into the top five.

What's your opinion on Terry Sawchuk?
Using your "SNW" (which I assume you are loosely referring to) he comes out even worse than Osgood does.
Now we know it's a fact that goaltenders on teams that focus on reducing shots are more likely to post a lower sv%. There are plenty of examples of this. Dominik Hasek in Detroit is an excellent example relevant to this; he was behind the same/similar defense and his numbers declined drastically from a Vezina winning season that was one of his best years. Did Hasek suddenly turn into an average goalie when he was traded? His SNW  and therefore his save percentage?  was below the league average multiple times in four seasons with the Red Wings and it was four of his five worst seasons. Yet his Ottawa numbers were better than his 2001 Vezina.
Why is that? Why would such a great goalie as Hasek have such a drop in statistics on one team?
And there's the Sawchuk thing, of course, where his overall number is only just slightly higher than average and his playoff number is actually well below.
It's interesting. You claim that save percentage is "basically" an indicator, and suggest that arguments against it are only worth a small margin of error, and then you use a set of advanced statistics that puts Osgood on the level of Terry Sawchuk, which is contrary to your original argument.
I disagree with your statistics, but I also disagree with your initial premise. You claim to be an expert on goaltending, and yet you are more interested in stats which do not account for a) unreliable shot counting and b) varying defensive systems which allow different amounts of scoring chances pershot. I don't agree with that method, and I think it's a poor way of evaluating talent because there are so many things that can happen which make situations unequal between goalies. A goaltender who faces fewer shots will naturally see a larger impact on his statistics when he does give up a goal. If his team makes a mistake that costs them a goal, that's the kind of thing stats don't track.